Remote sensing image registration method, device, electronic device and storage medium

Abstract

The present invention provides a remote sensing image registration method, apparatus, electronic device, and storage medium. The remote sensing image registration method includes: obtaining target feature points of target images, wherein the target images include a remote sensing image to be registered and a reference remote sensing image; for each target image, determining the overall phase consistency feature of the target image based on the local energy and phase intensity of the target image at all scales and directions; determining a logarithmic polar coordinate descriptor of the target feature points based on the overall phase consistency feature; and registering the remote sensing image to be registered and the reference remote sensing image based on the logarithmic polar coordinate descriptor of the target feature points. This solution can more accurately perform image registration of remote sensing images to be registered.

Description

Technical Field

[0001] The present invention relates to the technical field of remote sensing image processing, and in particular to a remote sensing image registration method, a remote sensing image registration device, an electronic device, a storage medium and a computer program product. Background Art

[0002] Image registration is the process of matching and superimposing two or more remote sensing images acquired at different times, using different sensors (imaging devices), or under different conditions (weather, illumination, camera position and angle, etc.). It has been widely used in the analysis of remote sensing data.

[0003] After extracting feature points from two or more remote sensing images, image registration can be performed on the two or more remote sensing images based on the feature point information corresponding to the feature points. The feature point information corresponding to the feature points will affect the effect of image registration.

[0004] For two or more remote sensing images with different conditions, traditional image registration methods typically use corresponding feature point extraction methods to extract feature points from the two or more remote sensing images, thereby improving the image registration effect. However, during image registration, the feature point information of the extracted feature points depends on the feature point extraction method. Because the image information in different remote sensing images varies, traditional image registration methods are not suitable for remote sensing image registration in complex situations. Summary of the Invention

[0005] The present invention has been made in view of the above-mentioned problems.

[0006] According to a first aspect of the present invention, a remote sensing image registration method is provided. The method comprises: obtaining target feature points of a target image, wherein the target image comprises a remote sensing image to be registered and a reference remote sensing image; for each of the target images, determining an overall phase consistency feature of the target image based on the local energy and phase intensity of the target image at all scales and directions; determining a logarithmic polar coordinate descriptor of the target feature point based on the overall phase consistency feature; and registering the remote sensing image to be registered with the reference remote sensing image based on the logarithmic polar coordinate descriptor of the target feature point.

[0007] Exemplarily, acquiring the target feature points of the target image includes: for each of the target images, determining a minimum distance and a maximum distance based on the phase intensity response of the target image; extracting a first corner feature point from the target image based on the minimum distance; extracting a first edge feature point from the target image based on the maximum distance; and determining the target feature point of the target image among the first edge feature points and the first corner feature points.

[0008] Exemplarily, determining the target feature points of the target image from the first edge feature points and the first corner feature points includes: obtaining the confidence of each first edge feature point and the confidence of each first corner feature point; dividing the target image into multiple regions; for each region, determining the first feature points and the second feature points in the region respectively, wherein the first feature points include the first edge feature points with the highest confidence in the region, and the distance between the first feature points in the region is not less than a first distance threshold, and the second feature points include the first corner feature points with the highest confidence in the region, and the distance between the second feature points in the region is not less than the first distance threshold; and selecting the target feature points of the target image from the first feature points and the second feature points.

[0009] Exemplarily, extracting a first corner feature point from the target image according to the minimum distance includes: using a FAST feature extractor to determine a second corner feature point according to the minimum distance; using a SIFT feature extractor to determine a third corner feature point according to the minimum distance; determining the same point between the second corner feature point and the third corner feature point as the first corner feature point; and / or

[0010] Extracting the first edge feature point from the target image according to the maximum distance includes: using a FAST feature extractor to determine a second edge feature point according to the maximum distance; using a SIFT feature extractor to determine a third edge feature point according to the maximum distance; and determining the same point between the second edge feature point and the third edge feature point as the first edge feature point.

[0011] Exemplarily, determining the log-polar coordinate descriptor of the target feature point based on the overall phase consistency feature includes: performing a convolution operation on the target image using an odd filter to obtain the overall phase direction feature of the target image; determining the first phase consistency feature of the field of each target feature point based on the position of the field of each target feature point and the overall phase consistency feature, wherein, for each target feature point, the pixel where the target feature point is located is the center pixel of the field of the target feature point; determining the first phase direction feature of the field of each target feature point based on the position of the field of each target feature point and the overall phase direction feature; for each target feature point, determining the log-polar coordinate descriptor of the target feature point based on the first phase consistency feature and the first phase direction feature of the field of the target feature point.

[0012] Exemplarily, the method of determining the log-polar coordinate descriptor of the target feature point based on the first phase consistency feature and the first phase direction feature of the domain of the target feature point includes: dividing the domain of the target feature point into multiple sub-regions; for each sub-region, determining the local direction and phase consistency strength information corresponding to each pixel of the sub-region based on the first phase consistency feature and the first phase direction feature of the domain of the target feature point; and determining the log-polar coordinate descriptor of the target feature point based on the local direction and phase consistency strength information corresponding to each pixel in each sub-region.

[0013] Exemplarily, the multiple sub-regions have equal areas.

[0014] Exemplarily, the method of determining the logarithmic polar coordinate descriptor of the target feature point based on the local direction and phase consistency strength information corresponding to each pixel in each sub-region includes: for the pixels in each sub-region, determining the pixels in the sub-region as pixels corresponding to different local direction ranges according to the local direction corresponding to each pixel; for each local direction range, adding the phase consistency strength information of the pixels corresponding to the local direction range to obtain the second phase consistency feature of the pixels corresponding to the local direction range; and determining the logarithmic polar coordinate descriptor of the target feature point based on the second phase consistency feature of the pixels in each sub-region corresponding to different local direction ranges.

[0015] According to a second aspect of the present invention, there is also provided a remote sensing image registration device, comprising:

[0016] A feature point acquisition module is used to obtain target feature points of a target image, wherein the target image includes a remote sensing image to be registered and a reference remote sensing image;

[0017] a feature determination module, configured to determine, for each of the target images, an overall phase consistency feature of the target image based on the local energy and phase intensity of the target image at all scales and directions;

[0018] A description module, configured to determine a logarithmic polar coordinate descriptor of a target feature point based on the overall phase consistency feature;

[0019] The registration module is used to register the remote sensing image to be registered and the reference remote sensing image according to the logarithmic polar coordinate descriptor of the target feature point.

[0020] According to a third aspect of the present invention, an electronic device is further provided, comprising: a processor and a memory, wherein the memory stores computer program instructions, and the computer program instructions are used by the processor to execute the above-mentioned remote sensing image registration method when the processor is running.

[0021] According to a fourth aspect of the present invention, a storage medium is further provided, on which program instructions are stored. When the program instructions are run, they are used to execute the above-mentioned remote sensing image registration method.

[0022] According to a fifth aspect of the present invention, a computer program product is further provided, comprising computer program instructions, wherein the computer program instructions are used to execute the above-mentioned remote sensing image registration method when running.

[0023] In the above technical solution, the target feature points of the target image are obtained, and the target image includes the remote sensing image to be registered and the reference remote sensing image. Then, for each of the target images, the overall phase consistency feature of the target image is determined based on the local energy and phase intensity of all scales and directions of the target image. Then, based on the overall phase consistency feature, the log polar coordinate descriptor of the target feature point is determined. Finally, based on the log polar coordinate descriptor of the target feature point, the remote sensing image to be registered and the reference remote sensing image are registered. The overall phase consistency feature of the target image enhances the response of the phase consistency information. The log polar coordinate descriptor of the target feature point determined based on the overall phase consistency feature can more accurately characterize the feature point information of the target feature point, so that the image registration of the remote sensing images to be registered can be performed more accurately.

[0024] The above description is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are specifically listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] The above and other objects, features, and advantages of the present invention will become more apparent through a more detailed description of the embodiments of the present invention with reference to the accompanying drawings. The accompanying drawings are provided to provide a further understanding of the embodiments of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and are not intended to limit the present invention. In the drawings, the same reference numerals generally represent the same components or steps.

[0026] Figure 1 A schematic flow chart of a remote sensing image registration method according to an embodiment of the present invention is shown;

[0027] Figure 2 A schematic flow chart of a method for acquiring target feature points of a target image according to an embodiment of the present invention is shown;

[0028] Figure 3 A schematic flowchart of determining target feature points of a target image from first edge feature points and first corner feature points according to an embodiment of the present invention is shown;

[0029] Figure 4 A schematic flow chart of extracting a first corner feature point from the target image according to a minimum distance according to one embodiment of the present invention is shown;

[0030] Figure 5 A schematic flowchart of extracting a first edge feature point from the target image according to the maximum distance according to one embodiment of the present invention is shown;

[0031] Figure 6 A schematic flow chart of determining a log-polar coordinate descriptor of a target feature point based on overall phase consistency features according to one embodiment of the present invention is shown;

[0032] Figure 7 A schematic flowchart of determining a log-polar coordinate descriptor of a target feature point based on a first phase consistency feature and a first phase direction feature of a field of the target feature point according to one embodiment of the present invention is shown;

[0033] Figure 8 A schematic flowchart of determining a log-polar coordinate descriptor of a target feature point based on local direction and phase consistency strength information corresponding to each pixel in each sub-region according to one embodiment of the present invention is shown;

[0034] Figure 9 A schematic diagram of determining a log-polar coordinate descriptor of a target feature point according to an embodiment of the present invention is shown;

[0035] Figure 10 A schematic block diagram of a remote sensing image registration device according to an embodiment of the present invention is shown;

[0036] Figure 11 A schematic block diagram of an electronic device according to an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0037] In order to make the purpose, technical solutions and advantages of the present invention more apparent, exemplary embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments of the present invention, and it should be understood that the present invention is not limited to the exemplary embodiments described herein. Based on the embodiments of the present invention described in the present invention, all other embodiments obtained by those skilled in the art without creative work should fall within the scope of protection of the present invention.

[0038] To at least partially address the above-mentioned issues, a remote sensing image registration method is proposed. The method obtains target feature points of a target image, wherein the target image includes a remote sensing image to be registered and a reference remote sensing image. Then, for each target image, the overall phase consistency feature of the target image is determined based on the local energy and phase intensity of the target image at all scales and directions. The logarithmic polar coordinate descriptor of the target feature point is then determined based on the overall phase consistency feature. Finally, the remote sensing image to be registered and the reference remote sensing image are registered based on the logarithmic polar coordinate descriptor of the target feature point. The overall phase consistency feature of the target image enhances the response of the phase consistency information. Determining the logarithmic polar coordinate descriptor of the target feature point based on the overall phase consistency feature can more accurately characterize the feature point information of the target feature point, thereby enabling more accurate image registration of the remote sensing images to be registered.

[0039] Figure 1 FIG. 1 shows a schematic flow chart of a remote sensing image registration method according to an embodiment of the present invention. Figure 1 As shown, the remote sensing image registration method may include steps S110 to S140.

[0040] In step S110 , target feature points of a target image are acquired, where the target image includes a remote sensing image to be registered and a reference remote sensing image.

[0041] The target feature points of the target image may include multiple types of feature points, such as corner points, spots, edge feature points, corner feature points, etc.

[0042] A feature extraction algorithm can be used to detect the remote sensing image to be registered and the reference remote sensing image respectively to obtain target feature points in the remote sensing image to be registered and the reference remote sensing image respectively. Alternatively, after obtaining the target feature points, the target feature points can be filtered and updated to obtain the desired target feature points in the remote sensing image to be registered and the reference remote sensing image respectively.

[0043] Step S120 : for each target image, determine the overall phase consistency feature of the target image according to the local energy and phase intensity of all scales and directions of the target image.

[0044] Each of the target images can be any suitable remote sensing image acquired using infrared remote sensing equipment. For example, the target image can be an image in the form of an RGB image, a grayscale image, a binary image, etc. The target image can be a static image or any video frame in a dynamic video. The target image can be an image of any suitable size and resolution. The target image can be an original image directly acquired by the remote sensing sensor, or an image after a preprocessing operation has been performed on the original image. The preprocessing operation can include all operations to improve the visual effect of the target image, increase its clarity, or highlight certain features in the image. For example, but not limitation, the preprocessing operation can include operations such as digitization, geometric transformation, normalization, and filtering of the original image. The target image can also be a synthesized image without affecting subsequent image processing. Among them, the baseline remote sensing image serves as the reference image, and the remote sensing image to be registered serves as the image that needs to be registered.

[0045] The target image can be analyzed at multiple scales using filters of varying sizes or a pyramid representation. Directional filters (such as Gabor filters) can be used to capture information in different directions. For each scale and direction combination, the local energy can be calculated separately. The local energy at a single scale and direction combination does not adequately reflect the texture characteristics of the entire target image, so it is necessary to calculate the local energy at multiple scales and directions of the target image. The number of directions and scales is not limited.

[0046] The target image can be analyzed using techniques such as Gabor filter or multi-scale wavelet transform to capture information of different scales and directions, and the corresponding phase intensity can be calculated to obtain the phase intensity of the target image at multiple scales and directions.

[0047] After obtaining local energies and phase intensities at multiple scales and directions, the local energies at all scales and directions can be summed to obtain a first summation result, and the phase intensities at all scales and directions can be summed to obtain a second summation result. Based on the first and second summation results, the overall phase congruency characteristics of the target image are then determined. The overall phase congruency characteristics thus obtained enhance the overall response of the phase congruency information and better reflect the complete image information of the target image.

[0048] For example, the overall phase consistency feature of the target image can be determined according to the following formula 1:

[0049]

[0050] Among them, PC FA Represents the overall phase consistency feature of the target image, ∑ s ∑ oEnergy s,o represents the sum of local energies of all scales and directions (the first addition result), ∑ s ∑ o An s,o It represents the sum of the phase intensities of all scales and directions (the second addition result), where s is the scale, o is the direction, and ξ is the bias term.

[0051] For example, the above-mentioned local energy may be weighted and noise-compensated local energy.

[0052] Step S130: Determine the log-polar coordinate descriptor of the target feature point based on the overall phase consistency feature.

[0053] The overall phase consistency feature and other relevant information (such as gradient direction) can be used to generate a log-polar coordinate descriptor for each target feature point. This descriptor can characterize the properties of each target feature point, thereby distinguishing each target feature point. It is understandable that different target feature points will have different properties due to their own position, the texture of the surrounding image area, and other image information. The log-polar coordinate descriptor can accurately distinguish each target feature point.

[0054] Step S140 , registering the remote sensing image to be registered with the reference remote sensing image according to the log-polar coordinate descriptor of the target feature point.

[0055] The log-polar descriptors of each target feature point in the remote sensing image to be registered and the reference remote sensing image can be compared. When the similarity between a log-polar descriptor in the remote sensing image to be registered and a log-polar descriptor in the reference remote sensing image meets a certain requirement, the two target feature points correspond. The remote sensing image to be registered and the reference remote sensing image can be registered based on each pair of corresponding target feature points.

[0056] In the above technical solution, target feature points of a target image are obtained, and the target image includes a remote sensing image to be registered and a reference remote sensing image. Then, for each target image, the overall phase consistency feature of the target image is determined based on the local energy and phase intensity of the target image at all scales and directions. Then, based on the overall phase consistency feature, the logarithmic polar coordinate descriptor of the target feature point is determined. Finally, based on the logarithmic polar coordinate descriptor of the target feature point, the remote sensing image to be registered and the reference remote sensing image are registered. The overall phase consistency feature of the target image enhances the response of the phase consistency information. Determining the logarithmic polar coordinate descriptor of the target feature point based on the overall phase consistency feature can more accurately characterize the feature point information of the target feature point, thereby enabling more accurate image registration of the remote sensing images to be registered.

[0057] Figure 2FIG. 1 is a schematic flow chart of a method for obtaining target feature points of a target image according to an embodiment of the present invention. Figure 2 As shown, the above step S110 may include steps S210 to S240.

[0058] In step S210 , for each of the target images, a minimum distance and a maximum distance are determined based on the phase intensity response of the target image.

[0059] The phase intensity responses of the target image in different directions and scales can be extracted by using Gabor filters or multi-scale wavelet transform.

[0060] The following explanation uses the Gabor filter as an example. A Fourier transform is performed on the target image to extract features, obtaining a frequency domain image of the target image. After setting the direction and scale, the Gabor filter can be used to perform a 2D Log-Gabor filter on the frequency domain image to obtain odd-symmetric and even-symmetric components at different directions and scales in the frequency domain. The odd-symmetric and even-symmetric components at different directions and scales in the frequency domain are then inverse Fourier transformed to obtain odd-symmetric and even-symmetric components in the spatial domain. Finally, the phase intensity response (intensity component) can be calculated based on the odd-symmetric and even-symmetric components in the spatial domain.

[0061] For example, the phase intensity response can be calculated using the following formula 2:

[0062]

[0063] Among them, An s,o is the phase intensity response (intensity component), s is the scale, o is the direction, E s,o is the even symmetric component in the spatial domain, O s,o is the odd symmetric component in the spatial domain.

[0064] After obtaining the phase intensity response (intensity component), the phase congruency information in each direction can be calculated based on the intensity component, the directionally weighted and noise-compensated local energy. Then, the phase congruency moments can be calculated based on the phase congruency information to obtain the maximum and minimum moments of the target image.

[0065] For example, the phase consistency information in each direction may be determined according to the following formula 3:

[0066]

[0067] Where PC(θ) is the phase consistency information in each direction, Energy(θ)s is the weighted and noise-compensated local energy in each direction, ∑ s An s,o(θ) is the sum of the phase intensity response (intensity components) in all directions.

[0068] For example, after obtaining the phase consistency information in each direction, the maximum moment and the minimum moment can be determined according to the following formulas 3 to 7:

[0069] a=∑ o (PC(θ)cos(θ)) 2 Formula 3

[0070] b=2∑ o (PC(θ)cos(θ))×(PC(θ)sin(θ)) Formula 4

[0071] c=∑ o (PC(θ)sin(θ)) 2 Formula 5

[0072]

[0073] Where a, b, and c are the intermediate covariance data in the phase consistency moment calculation, M max and M min are the maximum moment and the minimum moment respectively.

[0074] In step S220, a first corner feature point is extracted from the target image according to the minimum distance.

[0075] Any feature extraction algorithm may be used to extract the first corner feature point from the target image based on the minimum distance. The feature extraction algorithm may include FAST, SUPER, SIFT, etc.

[0076] In step S230 , a first edge feature point is extracted from the target image according to the maximum distance.

[0077] The first edge feature point can be extracted from the target image based on the minimum distance using any feature extraction algorithm, wherein the feature extraction algorithm may include FAST, SUPER, SIFT, etc.

[0078] The feature extraction algorithms used to extract the first corner feature points and the first edge feature points can be the same or different. For example, the first corner feature points can be extracted using the FAST feature extractor, and the first edge feature points can be extracted using the SUPER feature extractor. Another example is that the first edge feature points and the first corner feature points can both be extracted using the FAST feature extractor.

[0079] In step S240 , target feature points of the target image are determined from the first edge feature points and the first corner feature points.

[0080] The extracted first edge feature points and first corner feature points may not meet expectations and need to be screened again. For example, if the distance, confidence level, or number of the extracted first edge feature points and first corner feature points do not meet expectations, the first edge feature points and first corner feature points can be screened based on the corresponding expected conditions to obtain feature points that meet the expectations.

[0081] In the above technical solution, for each target image, the minimum distance and maximum distance are determined based on the phase intensity response of the target image. Then, based on the minimum distance and minimum moment, the first corner feature point and the first edge feature point are extracted from the target image, respectively. Finally, the target feature point of the target image is determined from the first edge feature points and the first corner feature points. This allows the edge feature points and the corner feature points to be combined with the phase congruence information of the target image to obtain the target feature point representing the joint edge and corner features, thereby improving the quality of the target feature point.

[0082] Figure 3 FIG. 1 shows a schematic flow chart of determining the target feature points of the target image from the first edge feature points and the first corner feature points according to an embodiment of the present invention. Figure 3 As shown, the above step S240 may include steps S310 to S340.

[0083] In step S310 , the confidence level of each first edge feature point and the confidence level of each first corner feature point are obtained.

[0084] A feature point detection algorithm can be used to perform feature point detection on the target image to determine the initial feature points of the target image and the confidence of each initial feature point, that is, when extracting each first edge feature point and each first corner feature point, the confidence of each first edge feature point and the confidence of each first corner feature point are determined respectively.

[0085] In step S320, the target image is divided into a plurality of regions.

[0086] The multiple regions can be of the same size or of different sizes. They can be of fixed size, or the size of each region can be determined based on the density of the first edge feature points and first corner feature points in the target image. The greater the density, the smaller the region; the smaller the density, the larger the region.

[0087] In step S330, for each region, the first feature point and the second feature point in the region are determined respectively, wherein the first feature point includes the first edge feature point with the highest confidence in the region, and the distance between the first feature points in the region is not less than the first distance threshold; the second feature point includes the first corner feature point with the highest confidence in the region, and the distance between the second feature points in the region is not less than the first distance threshold.

[0088] Optionally, the confidence levels between every two first edge feature points in the region can be compared, and the smaller of the two first edge feature points whose distance is less than a first distance threshold is removed until the distance between any two first edge feature points in the region is no less than the first distance threshold. The remaining first edge feature points are then used as first feature points. Similar operations are performed on the first corner feature points as for the first edge feature, resulting in the remaining first corner feature points being used as second feature points.

[0089] Optionally, starting with the first edge feature point with the highest confidence, all first edge feature points whose distance to the first edge feature point is less than a first distance threshold can be removed. This process can then be repeated, starting with the remaining first edge feature points with the highest confidence, until the distance between any two first edge feature points is no less than the first distance threshold. The remaining first edge feature points are then used as first feature points. Similar operations are performed on the first corner feature points as for the first edge feature, resulting in the remaining first corner feature points being used as second feature points.

[0090] In step S340 , a target feature point of the target image is selected from the first feature point and the second feature point.

[0091] For example, a confidence threshold may be set, first feature points and second feature points with confidences less than the confidence threshold may be removed, and the remaining first feature points and second feature points may be used as target feature points of the target image.

[0092] For example, a quantity threshold may be set to retain only the first feature points and second feature points with the highest confidence level not less than the quantity threshold, and then use the retained first feature points and second feature points as target feature points of the target image.

[0093] For example, the confidence of each first feature point and the second feature point can be normalized to obtain a normalized confidence corresponding to each first feature point and the second feature point. The first feature points and the second feature points whose confidence is less than the normalized confidence threshold are then removed, and the remaining first feature points and the second feature points are used as the target feature points of the target image.

[0094] In the above technical solution, the confidence of each first edge feature point and the confidence of each first corner feature point are obtained, and the target image is divided into multiple regions. Then, for each region, the first feature point and the second feature point in the region are determined respectively, wherein the first feature point includes the first edge feature point with the highest confidence in the region, and the distance between the first feature points in the region is not less than the first distance threshold, and the second feature point includes the first corner feature point with the highest confidence in the region, and the distance between the second feature points in the region is not less than the first distance threshold. Finally, the target feature point of the target image is selected from the first feature points and the second feature points. In this way, the obtained feature points can be screened to obtain high-quality target feature points.

[0095] Figure 4 FIG. 1 shows a schematic flow chart of extracting the first corner feature point from the target image according to the minimum distance according to an embodiment of the present invention. Figure 4 As shown, the above step S220 may include steps S410 to S430.

[0096] In step S410, the FAST feature extractor is used to determine the second corner feature point according to the minimum distance.

[0097] In step S420, a SIFT feature extractor is used to determine a third corner feature point according to the minimum distance.

[0098] In step S430, the same point in the second corner feature point and the third corner feature point is determined as the first corner feature point.

[0099] The FAST feature extractor is the FAST feature point detection algorithm, and the SIFT feature extractor is the SIFT feature point detection algorithm. Using different feature point detection algorithms, we can obtain second-corner and third-corner feature points with different characteristics. By identifying the common points among the second-corner and third-corner feature points as first-corner feature points, we can combine the advantages of both feature point detection algorithms to obtain more effective first-corner feature points.

[0100] In the above technical solution, the FAST feature extractor is used to determine the second corner feature point based on the minimum distance, and the SIFT feature extractor is used to determine the third corner feature point based on the minimum distance. Then, the points that are identical between the second and third corner feature points are determined as the first corner feature point. This can extract the first corner feature point with better results.

[0101] Figure 5 FIG. 5 shows a schematic flow chart of extracting the first edge feature point from the target image according to the maximum distance according to an embodiment of the present invention. Figure 5 As shown, the above step S230 may include steps S510 to S530.

[0102] In step S510, the FAST feature extractor is used to determine the second edge feature point according to the maximum distance.

[0103] In step S520 , a SIFT feature extractor is used to determine a third edge feature point according to the maximum distance.

[0104] In step S530 , the same point between the second edge feature point and the third edge feature point is determined as the first edge feature point.

[0105] Similar to steps S410 to S420 above, different feature point detection algorithms can be used to obtain second and third edge feature points with different characteristics. The same points among the second and third edge feature points are then identified as first edge feature points. This combines the advantages of the two feature point detection algorithms to obtain a more effective first edge feature point. This will not be discussed in detail here.

[0106] In the above technical solution, the FAST feature extractor is used to determine the second edge feature point based on the minimum distance, and the SIFT feature extractor is used to determine the third edge feature point based on the minimum distance. Then, the points that are identical between the second and third edge feature points are determined as the first edge feature point. In this way, the first edge feature point can be extracted with better effect.

[0107] Optionally, the SIFT feature extractor for extracting the third corner feature points and the third edge feature points may be replaced by other feature extractors, such as SURF, ORB, etc.

[0108] For example, the above steps S410 to S430 and the above steps S510 to S530 may be combined to obtain first edge feature points and first corner feature points with better effects.

[0109] Figure 6 FIG. 1 shows a schematic flow chart of determining the logarithmic polar coordinate descriptor of a target feature point based on the overall phase consistency feature according to an embodiment of the present invention. Figure 6 As shown, the above step S130 may include steps S610 to S630.

[0110] In step S610, a convolution operation is performed on the target image using an odd-numbered filter to obtain an overall phase direction feature of the target image.

[0111] You can first set the direction of an odd filter (such as a Gabor filter), and then use the odd filter to perform a convolution operation on the target image to obtain convolution results in different directions. Then, based on the phase direction characteristics of each direction and scale of the target image, you can obtain the overall phase direction characteristics of the target image.

[0112] For example, the overall phase direction feature of the target image can be determined according to the following formula 8.

[0113]

[0114] Among them, Sum odd (θ) represents the sum of the convolution results obtained by the odd filter operation under the direction θ, and PO represents the overall phase direction feature. The overall phase direction feature of the target image represents the angular information of the local direction corresponding to all pixels of the target image.

[0115] For example, when PO is less than 0, the forward angle reversal caused by convolution is eliminated by adding π.

[0116] In step S620, the first phase consistency feature of the field of each target feature point is determined based on the position of the field of each target feature point and the overall phase consistency feature, wherein, for each target feature point, the pixel where the target feature point is located is the center pixel of the field of the target feature point.

[0117] The domain of the target feature point can be in the shape of a regular shape such as a circle or a rectangle to facilitate calculation. Among them, when the domain of the target feature point is a circle, it is more conducive to determining the logarithmic polar coordinate descriptor.

[0118] The area of ​​the target feature point may be a preset size, for example, a circle with a diameter of 33 pixels.

[0119] The location of each target feature point's domain can be determined based on the edge coordinates of its domain. For example, if the target feature point's domain is a circle with a diameter of 33 pixels, the leftmost edge coordinates are (a, b), the top edge coordinates are (c, d), the rightmost edge coordinates are (a+32, b), and the bottom edge coordinates are (c, d-32). Based on these four coordinates, the location of the circular target feature point's domain and the pixels within that domain can be determined.

[0120] The overall phase consistency feature represents the phase consistency information of each feature point of the target image. Therefore, according to the position of the area of ​​each target feature point, the local phase consistency feature corresponding to the pixels in each area can be determined as the first phase consistency feature.

[0121] In step S630 , a first phase direction feature of the area of ​​each target feature point is determined according to the position of the area of ​​each target feature point and the overall phase direction feature.

[0122] Similar to step S640, it will not be described in detail here.

[0123] In step S640 , for each target feature point, a log-polar coordinate descriptor of the target feature point is determined based on the first phase consistency feature and the first phase direction feature of the domain of the target feature point.

[0124] The image of the domain of each target feature point can be mapped to the polar index space, and the logarithmic polar coordinate descriptor of the target feature point can be determined based on the pixels, first phase consistency features and first phase direction features in the domain of the target feature point in the polar index space.

[0125] For example, the logarithmic polar coordinate descriptor can be data in the form of a vector, and the number of elements of the vector can be determined according to the number of local directions represented by the first phase direction feature and the number of sub-regions in the field of the target feature point. The data in each element represents the sum of the intensities of the phase consistency information (first phase consistency feature) of each pixel in the same local direction in the corresponding sub-region. The intensity of the phase consistency information represents the phase consistency intensity information.

[0126] In the above technical solution, an odd-numbered filter is used to perform a convolution operation on the target image to obtain the overall phase direction feature of the target image, and then the first phase consistency feature of the field of each target feature point is determined according to the position of the field of each target feature point and the overall phase consistency feature, wherein, for each target feature point, the pixel where the target feature point is located is the center pixel of the field of the target feature point, and then the first phase direction feature of the field of each target feature point is determined according to the position of the field of each target feature point and the overall phase direction feature, and finally, for each target feature point, the logarithmic polar coordinate descriptor of the target feature point is determined according to the first phase consistency feature and the first phase direction feature of the field of the target feature point. In this way, the overall phase consistency feature and the overall phase direction feature can be combined to determine a descriptor that can accurately describe the target feature point.

[0127] Figure 7 FIG. 1 shows a schematic flow chart of determining the logarithmic polar coordinate descriptor of the target feature point according to the first phase consistency feature and the first phase direction feature of the field of the target feature point according to an embodiment of the present invention. Figure 7 As shown, the above step S130 may include steps S710 to S730.

[0128] In step S710 , the area of ​​the target feature point is divided into a plurality of sub-areas.

[0129] Exemplarily, the areas of the multiple sub-regions are equal.

[0130] When the areas of multiple sub-regions are equal, calculation errors caused by different amounts of image information contained in each sub-region can be avoided.

[0131] In step S720, for each sub-region, local direction and phase consistency strength information corresponding to each pixel in the sub-region is determined according to the first phase consistency feature and the first phase direction feature of the domain of the target feature point.

[0132] For each sub-region, the pixels in the sub-region can be determined according to the position of the sub-region, and the local direction and phase consistency strength information corresponding to each pixel in the sub-region can be determined. Here, reference can be made to the above-mentioned step S620, which will not be described in detail here.

[0133] In step S730, a log-polar coordinate descriptor of the target feature point is determined based on the local direction and phase consistency strength information corresponding to each pixel in each sub-region.

[0134] The image of each sub-region can be mapped to the polar index space, and the logarithmic polar coordinate descriptor of the target feature point can be determined based on the pixels in the sub-region under the polar index space, the local direction corresponding to each pixel, and the phase consistency strength information.

[0135] For example, the log-polar coordinate descriptor can be data in the form of a vector, and the number of elements in the vector can be determined based on the number of local directions corresponding to each pixel in the subregion and the total number of subregions. The data in each element represents the intensity of phase congruency information for each pixel in the corresponding subregion in the same local direction.

[0136] In the above technical solution, the domain of the target feature point is divided into multiple sub-regions. Then, for each sub-region, the local direction and phase consistency strength information corresponding to each pixel in the sub-region are determined based on the first phase consistency feature and the first phase direction feature of the domain of the target feature point. Finally, the logarithmic polar coordinate descriptor of the target feature point is determined based on the local direction and phase consistency strength information corresponding to each pixel in each sub-region. This can refine the domain of the target feature point and obtain a more accurate logarithmic polar coordinate descriptor, which is more conducive to image registration.

[0137] Figure 8 FIG1 shows a schematic flow chart of determining the logarithmic polar coordinate descriptor of the target feature point according to the local direction and phase consistency strength information corresponding to each pixel in each sub-region according to an embodiment of the present invention. Figure 8 As shown, the above step S730 may include steps S810 to S830.

[0138] In step S810 , for pixels in each sub-region, the pixels in the sub-region are determined as pixels corresponding to different local direction ranges according to the local directions corresponding to each pixel.

[0139] Each local direction range may include multiple local directions. For example, a local direction may be any direction within 360 degrees. Local directions may be divided into multiple continuous intervals as local direction ranges. The local direction corresponding to each pixel is then divided into its own local direction range. For example, if a local direction range exists between 0 and 45 degrees and a pixel corresponds to a local direction of 30 degrees, the pixel may be classified as a pixel within the local direction range of 0 to 45 degrees.

[0140] In step S820, for each local direction range, the phase consistency strength information of the pixels corresponding to the local direction range is added to obtain a second phase consistency feature of the pixels corresponding to the local direction range.

[0141] In step S830, a log-polar coordinate descriptor of the target feature point is determined based on the second phase-consistency features of pixels in each sub-region corresponding to different local direction ranges.

[0142] The log-polar coordinate descriptor can be represented by a vector. The second phase consistency features of pixels in different local direction ranges can be used as each element in the log-polar coordinate descriptor, and then the log-polar coordinate descriptor of the target feature point can be obtained.

[0143] For example, before determining each element in the log-polar coordinate descriptor, a weight coefficient may be determined based on the distance of the sub-region from the target feature point. The pixels in the same local direction within the sub-region that are closer to the target feature point have a higher weight coefficient. The second phase congruency feature weighted by the weight coefficient is used as the value of each element in the log-polar coordinate descriptor.

[0144] In the above technical solution, for pixels in each subregion, the pixels in the subregion are determined to correspond to different local direction ranges based on the local directions corresponding to each pixel. Then, for each local direction range, the phase consistency strength information of the pixels corresponding to the local direction range is added to obtain the second phase consistency feature of the pixels corresponding to the local direction range. Finally, based on the second phase consistency features of the pixels in each subregion corresponding to the different local direction ranges, the logarithmic polar coordinate descriptor of the target feature point is determined. This discretizes the local directions and reduces the amount of data required to determine the logarithmic polar coordinate descriptor.

[0145] Figure 9 FIG. 1 shows a schematic diagram of determining a logarithmic polar coordinate descriptor of a target feature point according to an embodiment of the present invention. Figure 9As shown, after determining the overall phase congruency feature and the overall phase orientation feature of the target image, the area of ​​the target feature point P(x, y) can be divided into 32 sub-regions of equal size, and the 8 local direction ranges corresponding to the pixels in each sub-region are determined. The local direction range to which each pixel belongs is determined based on the 8 local direction ranges. Then, based on the 32 sub-regions and the 8 local direction ranges, a 256-bit logarithmic polar coordinate descriptor is determined. The polar coordinate descriptor can be represented in vector form, with each element representing the sum of the intensity values ​​of the phase congruency information corresponding to the pixels in the same local direction range within the sub-region.

[0146] Figure 10 FIG. 1 shows a schematic block diagram of a remote sensing image registration device according to an embodiment of the present invention. Figure 10 As shown, the remote sensing image registration device includes a feature point acquisition module 1010 , a feature determination module 1020 , a description module 1030 , and a registration module 1040 .

[0147] The feature point acquisition module 1010 is used to obtain target feature points of a target image, where the target image includes a remote sensing image to be registered and a reference remote sensing image.

[0148] The feature determination module 1020 is configured to determine, for each target image, an overall phase consistency feature of the target image based on the local energy and phase intensity of the target image at all scales and directions.

[0149] The description module 1030 is used to determine the log-polar coordinate descriptor of the target feature point according to the overall phase consistency feature.

[0150] The registration module 1040 is used to register the remote sensing image to be registered with the reference remote sensing image according to the log-polar coordinate descriptor of the target feature points.

[0151] For example, the feature point acquisition module 1010 includes a first determination submodule, a first extraction submodule, a second extraction submodule, and a second determination submodule. The first determination submodule is configured to determine, for each target image, a minimum distance and a maximum distance based on the phase intensity response of the target image. The second extraction submodule is configured to extract a first corner feature point from the target image based on the minimum distance. The third extraction submodule is configured to extract a first edge feature point from the target image based on the maximum distance. The second determination submodule is configured to determine a target feature point of the target image from the first edge feature point and the first corner feature point.

[0152] Exemplarily, the second determination submodule includes a first acquisition submodule, a first division submodule, a third determination submodule and a first selection submodule. The first acquisition submodule is used to obtain the confidence of each first edge feature point and the confidence of each first corner feature point. The first division submodule is used to divide the target image into multiple regions. The third determination submodule is used to determine the first feature point and the second feature point in each region, respectively, wherein the first feature point includes the first edge feature point with the highest confidence in the region, and the distance between the first feature points in the region is not less than the first distance threshold, and the second feature point includes the first corner feature point with the highest confidence in the region, and the distance between the second feature points in the region is not less than the first distance threshold. The first selection submodule is used to select the target feature point of the target image from the first feature point and the second feature point.

[0153] Exemplarily, the first extraction submodule includes a first feature extraction submodule, a second feature extraction submodule, and a corner feature point determination submodule; and / or the second extraction submodule includes a third feature extraction submodule, a fourth feature extraction submodule, and an edge feature point determination submodule. The first feature extraction submodule is used to determine the second corner feature point based on the minimum distance using the FAST feature extractor. The second feature extraction submodule is used to determine the third corner feature point based on the minimum distance using the SIFT feature extractor. The corner feature point determination submodule is used to determine the same point in the second corner feature point and the third corner feature point as the first corner feature point. The third feature extraction submodule is used to determine the second edge feature point based on the maximum distance using the FAST feature extractor. The fourth feature extraction submodule is used to determine the third edge feature point based on the maximum distance using the SIFT feature extractor. The edge feature point determination submodule is used to determine the same point in the second edge feature point and the third edge feature point as the first edge feature point.

[0154] Exemplarily, the description module 1030 includes a convolution submodule, a first phase consistency feature determination submodule, a first phase direction feature determination submodule and a first descriptor determination submodule. The convolution submodule is used to perform a convolution operation on the target image using an odd filter to obtain the overall phase direction feature of the target image. The first phase consistency feature determination submodule is used to determine the first phase consistency feature of the field of each target feature point based on the position of the field of each target feature point and the overall phase consistency feature, wherein, for each target feature point, the pixel where the target feature point is located is the center pixel of the field of the target feature point. The first phase direction feature determination submodule is used to determine the first phase direction feature of the field of each target feature point based on the position of the field of each target feature point and the overall phase direction feature. The first descriptor determination submodule is used to determine the logarithmic polar coordinate descriptor of the target feature point for each target feature point based on the first phase consistency feature and the first phase direction feature of the field of the target feature point.

[0155] The first descriptor determination submodule includes a second division submodule, a pixel analysis submodule and a second descriptor determination submodule. The second division submodule is used to divide the domain of the target feature point into multiple subregions. The pixel analysis submodule is used to determine, for each subregion, the local direction and phase consistency strength information corresponding to each pixel in the subregion based on the first phase consistency feature and the first phase direction feature of the domain of the target feature point. The second descriptor determination submodule is used to determine the logarithmic polar coordinate descriptor of the target feature point based on the local direction and phase consistency strength information corresponding to each pixel in each subregion.

[0156] Exemplarily, the areas of the multiple sub-regions are equal.

[0157] Exemplarily, the second descriptor determination submodule includes a pixel allocation submodule, a second phase consistency feature determination submodule and a third descriptor determination submodule. The pixel allocation submodule is used to determine the pixels in each subregion as pixels corresponding to different local direction ranges according to the local directions corresponding to each pixel. The second phase consistency feature determination submodule is used to add the phase consistency intensity information of the pixels corresponding to the local direction range for each local direction range to obtain the second phase consistency feature of the pixels corresponding to the local direction range. The third descriptor determination submodule is used to determine the logarithmic polar coordinate descriptor of the target feature point based on the second phase consistency features of the pixels in each subregion corresponding to different local direction ranges.

[0158] According to another aspect of the present invention, an electronic device is provided. Figure 11 FIG. 1 shows a schematic block diagram of an electronic device according to an embodiment of the present invention. Figure 11 As shown, the electronic device includes a processor and a memory, wherein the memory stores computer program instructions, and the computer program instructions are used by the processor to execute the remote sensing image registration method as described above when the processor is running.

[0159] In addition, according to another aspect of the present invention, a storage medium is provided, on which program instructions are stored. When the program instructions are executed by a computer or processor, the computer or processor is caused to perform the corresponding steps of the remote sensing image registration method described above in the embodiment of the present invention, and is used to implement the corresponding modules in the remote sensing image registration device described above in the embodiment of the present invention. The storage medium may include, for example, a storage component of a tablet computer, a hard disk of a personal computer, a read-only memory (ROM), an erasable programmable read-only memory (EPROM), a portable compact disc read-only memory (CD-ROM), a USB memory, or any combination of the above storage media. The computer-readable storage medium may be any combination of one or more computer-readable storage media.

[0160] According to another aspect of the present invention, a computer program product is provided, comprising computer program instructions, wherein the computer program instructions are used to execute the above remote sensing image registration method when running.

[0161] A person skilled in the art can understand the specific implementation and beneficial effects of the above-mentioned remote sensing image registration device, electronic device, storage medium and computer program product by reading the above-mentioned detailed description of the remote sensing image registration method. For the sake of brevity, they will not be repeated here.

[0162] Although example embodiments have been described herein with reference to the accompanying drawings, it should be understood that the above example embodiments are merely illustrative and are not intended to limit the scope of the present invention. Various changes and modifications may be made therein by those skilled in the art without departing from the scope and spirit of the present invention. All such changes and modifications are intended to be included within the scope of the present invention as claimed in the appended claims.

[0163] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present invention.

[0164] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the units described is merely a logical function division. In actual implementation, other division methods may be used, such as combining or integrating multiple units or components into another device, or ignoring or not performing some features.

[0165] In the description provided herein, numerous specific details are described. However, it is understood that embodiments of the present invention may be practiced without these specific details. In some instances, well-known methods, structures, and techniques are not shown in detail so as not to obscure the understanding of this description.

[0166] Similarly, it should be understood that in order to streamline the present invention and aid in understanding one or more of the various inventive aspects, in the description of exemplary embodiments of the present invention, the various features of the present invention are sometimes grouped together into a single embodiment, figure, or description thereof. However, this approach to the present invention should not be interpreted as reflecting the intention that the claimed invention requires more features than those explicitly recited in each claim. More precisely, as reflected in the corresponding claims, the inventive point is that the corresponding technical problem can be solved with fewer features than all the features of a single disclosed embodiment. Therefore, the claims following the detailed description are hereby expressly incorporated into the detailed description, with each claim itself serving as a separate embodiment of the present invention.

[0167] It will be understood by those skilled in the art that, except where mutually exclusive, all features disclosed in this specification (including the accompanying claims, abstract, and drawings) and all processes or units of any method or apparatus disclosed herein may be combined in any combination. Unless expressly stated otherwise, each feature disclosed in this specification (including the accompanying claims, abstract, and drawings) may be replaced by an alternative feature providing the same, equivalent, or similar purpose.

[0168] Furthermore, those skilled in the art will appreciate that although some embodiments described herein include certain features included in other embodiments but not other features, combinations of features from different embodiments are intended to be within the scope of the present invention and to form different embodiments. For example, in the claims, any of the claimed embodiments may be used in any combination.

[0169] The various component embodiments of the present invention may be implemented in hardware, or in software modules running on one or more processors, or in a combination thereof. Those skilled in the art will appreciate that a microprocessor or digital signal processor (DSP) may be used in practice to implement some or all of the functions of some modules in the remote sensing image registration apparatus according to an embodiment of the present invention. The present invention may also be implemented as a device program (e.g., a computer program and a computer program product) for executing part or all of the methods described herein. Such a program for implementing the present invention may be stored on a computer-readable medium, or may be in the form of one or more signals. Such a signal may be downloaded from an Internet website, or provided on a carrier signal, or provided in any other form.

[0170] It should be noted that the above embodiments illustrate rather than limit the invention, and that those skilled in the art may devise alternative embodiments without departing from the scope of the appended claims. In the claims, any reference signs placed between brackets should not be construed as limiting the claims. The word "comprising" does not exclude the presence of elements or steps not listed in the claims. The word "a" or "an" preceding an element does not exclude the presence of a plurality of such elements. The present invention may be implemented by means of hardware comprising several different elements and by means of appropriately programmed computers. In a unit claim enumerating several means, several of these means may be embodied by the same item of hardware. The use of the words first, second, and third etc. does not indicate any order. These words may be interpreted as names.

[0171] The foregoing description is merely a specific embodiment of the present invention or an illustration of a specific embodiment. The scope of protection of the present invention is not limited thereto. Any modifications or substitutions that can be readily conceived by a person skilled in the art within the technical scope disclosed in the present invention are intended to be encompassed by the scope of protection of the present invention. The scope of protection of the present invention shall be subject to the scope of protection of the claims.

Claims

1. A remote sensing image registration method, characterized in that: The method comprises: Acquire target feature points of a target image, wherein the target image includes a remote sensing image to be registered and a reference remote sensing image; For each of the target images, determining an overall phase consistency feature of the target image based on the local energy and phase intensity of the target image at all scales and directions; Determining a logarithmic polar coordinate descriptor of a target feature point according to the overall phase consistency feature; Registering the remote sensing image to be registered and the reference remote sensing image according to the logarithmic polar coordinate descriptor of the target feature point; The step of determining the logarithmic polar coordinate descriptor of the target feature point based on the overall phase consistency feature includes: Performing a convolution operation on the target image using an odd-numbered filter to obtain an overall phase direction feature of the target image; Determine a first phase consistency feature of the domain of each target feature point according to the position of the domain of each target feature point and the overall phase consistency feature, wherein, for each target feature point, the pixel where the target feature point is located is the center pixel of the domain of the target feature point; Determining a first phase direction feature of the field of each target feature point according to the position of the field of each target feature point and the overall phase direction feature; For each target feature point, a log-polar coordinate descriptor of the target feature point is determined according to the first phase consistency feature and the first phase direction feature of the domain of the target feature point.

2. The method according to claim 1, characterized in that The acquiring target feature points of the target image includes: For each of the target images, Determine the minimum distance and the maximum distance according to the phase intensity response of the target image; extracting a first corner feature point from the target image according to the minimum distance; extracting a first edge feature point from the target image according to the maximum distance; Target feature points of the target image are determined from the first edge feature points and the first corner feature points.

3. The method according to claim 2, characterized in that Determining the target feature points of the target image from the first edge feature points and the first corner feature points includes: Obtaining the confidence of each first edge feature point and the confidence of each first corner feature point; Dividing the target image into multiple regions; For each region, determine the first feature point and the second feature point in the region respectively, wherein the first feature point includes the first edge feature point with the highest confidence in the region, and the distance between the first feature points in the region is not less than a first distance threshold; the second feature point includes the first corner feature point with the highest confidence in the region, and the distance between the second feature points in the region is not less than the first distance threshold; A target feature point of the target image is selected from the first feature point and the second feature point.

4. The method according to claim 2, characterized in that The extracting the first corner feature point from the target image according to the minimum distance includes: Using a FAST feature extractor, determining a second corner feature point according to the minimum distance; Using a SIFT feature extractor, determining a third corner feature point according to the minimum distance; determining the same point between the second corner feature point and the third corner feature point as the first corner feature point; and / or Extracting a first edge feature point from the target image according to the maximum distance includes: Determine a second edge feature point according to the maximum distance using a FAST feature extractor; Using a SIFT feature extractor, determining a third edge feature point according to the maximum distance; The same point between the second edge feature point and the third edge feature point is determined as the first edge feature point.

5. The method according to claim 1, wherein The step of determining the logarithmic polar coordinate descriptor of the target feature point according to the first phase consistency feature and the first phase direction feature of the field of the target feature point includes: Divide the area of ​​the target feature point into multiple sub-areas; For each sub-region, determining the local direction and phase consistency strength information corresponding to each pixel in the sub-region according to the first phase consistency feature and the first phase direction feature of the domain of the target feature point; According to the local direction and phase consistency strength information corresponding to each pixel in each sub-region, the logarithmic polar coordinate descriptor of the target feature point is determined.

6. The method according to claim 5, characterized in that The multiple sub-regions have equal areas.

7. The method according to claim 5, characterized in that Determining the logarithmic polar coordinate descriptor of the target feature point according to the local direction and phase consistency strength information corresponding to each pixel in each sub-region includes: For pixels in each sub-region, determining the pixels in the sub-region as pixels corresponding to different local direction ranges according to the local directions corresponding to each pixel; For each local direction range, summing the phase consistency strength information of the pixels corresponding to the local direction range to obtain a second phase consistency feature of the pixels corresponding to the local direction range; A logarithmic polar coordinate descriptor of the target feature point is determined according to the second phase consistency features of pixels in each sub-region corresponding to different local direction ranges.

8. A remote sensing image registration device, characterized in that: include: A feature point acquisition module is used to obtain target feature points of a target image, wherein the target image includes a remote sensing image to be registered and a reference remote sensing image; a feature determination module, configured to determine, for each of the target images, an overall phase consistency feature of the target image based on the local energy and phase intensity of the target image at all scales and directions; A description module, configured to determine a logarithmic polar coordinate descriptor of a target feature point based on the overall phase consistency feature; a registration module, configured to register the remote sensing image to be registered with the reference remote sensing image according to the logarithmic polar coordinate descriptor of the target feature point; The step of determining the logarithmic polar coordinate descriptor of the target feature point based on the overall phase consistency feature includes: Performing a convolution operation on the target image using an odd-numbered filter to obtain an overall phase direction feature of the target image; Determine a first phase consistency feature of the domain of each target feature point according to the position of the domain of each target feature point and the overall phase consistency feature, wherein, for each target feature point, the pixel where the target feature point is located is the center pixel of the domain of the target feature point; Determining a first phase direction feature of the field of each target feature point according to the position of the field of each target feature point and the overall phase direction feature; For each target feature point, a log-polar coordinate descriptor of the target feature point is determined according to the first phase consistency feature and the first phase direction feature of the domain of the target feature point.

9. An electronic device comprising a processor and a memory, characterized in that: The memory stores computer program instructions, which are used by the processor to execute the remote sensing image registration method according to any one of claims 1 to 7 when the processor is running the computer program instructions.

10. A storage medium having program instructions stored thereon, characterized in that: The program instructions are used to execute the remote sensing image registration method according to any one of claims 1 to 7 when running.

11. A computer program product comprising computer program instructions, characterized in that The computer program instructions are used to execute the remote sensing image registration method according to any one of claims 1 to 7 when running.

Images